Dalila Bouras

Dalila Bouras

Auto-generated profile, based on publicly available information

Dalila Bouras

Department of Mathematics and Computer Science, Larbi Tebessi University, Tebessa, Algeria, and Laboratory of Mathematics, Informatics, and Systems (LAMIS), Larbi Tebessi University, Tebessa, Algeria

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 June 2020 in Recent Advances in Computer Science and Communications
Verified
Journal Article
Closed access
PDF file
DBDalila Bouras
Dalila Bouras
MAMohamed Amroune
Mohamed Amroune
HBHakim Bendjenna
Hakim Bendjenna
NANabiha Azizi
Nabiha Azizi
Background: Nowadays, with the appearance of web 2.0, more users express their opinions, judgments, and thoughts towards certain objects, services, organizations, and their attributes via social networking, forum entries, websites, and blogs and so on. In this way, the volume of raw content generated by these users will increase rapidly with enormous size, where people often find difficulties in identifying and summarizing fine-grained sentiments buried in the opinion-rich resources. The traditional opinion mining techniques, which focused on the overall sentiment of the review, fails to uncover the sentiments expressed on the aspects of the reviewed entity. For that, researchers in Aspect-based opinion mining community try to solve and handle this problem. Objective: Our proposed study aims to present, survey and compare in the first place the important recent Aspect-based opinion mining approaches relevant to important languages such English, Arabic and Chinese and commonly datasets used in literature so that future researchers could improve their results. The cited approaches used the last techniques in the area on Opinion mining field, relevant to the Deep Learning models. In the second place, we try to highlight and give special attention to the Arabic language by introducing a dashboard of deep learning methods dedicated to the Arabic language. Finally, we emphasize the research gaps and future challenges in both English and Arabic languages that provide some new potential research fields. Methods: We have carefully summarized 48 models according to their algorithm into three categories: supervised, semi-supervised and unsupervised. Due to a large number of approaches with diverse datasets and techniques, we propose some statistical graphics to compare different experimentation results namely precision, Recall, and F-measure. Also, the study has conferred a comparative analysis and a comprehensive discussion of different approaches and techniques dedicated to the aspect extraction sub-task using the new tendency that of deep learning on both Arabic, English and Chinese language. We have introduced some future challenges, research gaps, and new trends in the opinion mining task, which need more efforts and investigations to present new solutions that make the opinion mining field more pervasive and give more ideas about the different researches done in the field of OM. Methods: We have compared the different approaches and techniques dedicated to the extraction of aspects using the new tendency that of deep learning. Our contribution illustrates the add values given by deep learning models in the treatment of user reviews expressed in the Arabic language. At the same time, this work is mainly based on the use of the evaluation performance metrics (precision, recall, and f-measure).

1 May 2022 in Recent Advances in Computer Science and Communications
Verified
Journal Article
Closed access
PDF file
DBDalila Bouras
Dalila Bouras
MAMohamed Amroune
Mohamed Amroune
HBHakim Bendjenna
Hakim Bendjenna
IBIssam Bendib
Issam Bendib
Objective: One key task of fine-grained opinion mining on product review is to extract product aspects and their corresponding opinion expressed by users. Previous work has demonstrated that precise modeling of opinion targets within the surrounding context can improve performances. However, how to effectively and efficiently learn hidden word semantics and better represent targets and the context still needs to be further studied. Recent years have seen a revival of the Long Short- Term Memory (LSTM), with its effectiveness being demonstrated on a wide range of problems. However, LSTM based approaches are still limited to linear data processing since it processes the information sequentially. As a result, they may perform poorly on user-generated texts, such as product reviews, tweets, etc., whose syntactic structure is not precise. Methods: In this research paper, we propose a constituency tree long short term memory neural network- based approach. We compare our model with state-of-the-art baselines on SemEval 2014 datasets. Results: Experiment results show that our models obtain competitive performances compared to various supervised LSTM architectures. Conclusion: Our work contributes to the improvement of state-of-the-art aspect-level opinion mining methods and offers a new approach to support the human decision-making process based on opinion mining results.